Jump to content

Clover General discussion

ErmaC

By ErmaC,
in Clover

 Share
30,171 posts in this topic

Recommended Posts

TheBloke
TheBloke
Posted
Posted

@TheBloke

Installing Clover you may choose SATA driver or BiosBlockIO.

 

In the first case you will see 6 on the screen. In the second case 7.

I may propose that for LSI controller you need 7.

 

Thanks very much for the fast reply.  I have tried this, but unfortunately it does not work.

 

I tried two things:

 

1. First I re-installed Clover to my USB stick, selecting "BiosBlockIO" instead of "SATA".  Booting with the USB stick gave me '7' in top left as you said, but then it did not list any drive connected to the LSI controller.  Same as before.

 

2. Then I tried re-installing Clover on my SSD, and booting direct from the LSI controller.  Again this showed '7', but then it failed to load the Clover GUI again.  It instead went to the BIOS-like screen with "Continue" option, meaning it couldn't load CloverX64.efi.   This is basically the same as when I first tried booting Clover direct from LSI.  The only difference is that this time, in the BIOS-like interface, one of the entries showed some data - the menu option "Boot Manager" showed two disks.  But pressing Enter on either of them just took me back to the Clover BIOS-like interface.

 

These are the settings I used when re-installing Clover on my SSD:

 

UFhpW6J.png

 

Mp2zv2D.png

 

Most of these were selected by default, I just changed SATA to BiosBlockIO.

 

If there is anything else left to try, I would be most grateful for any ideas.  Thanks again for your help.

Link to comment
Share on other sites
Matgen84
Matgen84
Posted
Posted

Hi,

 

I found that in preboot.log: I don't know why same DDR use different XMP version and Frequency?

=== [ ScanSPD ] ===========================================
1:806  0:000  SMBus device : 8086 1E22 class=0C0500 status=Success
1:806  0:000  SMBus CmdReg: 0x3
1:806  0:000  Scanning SMBus [8086:1E22], mmio: 0xF7135004, ioport: 0xF000, hostc: 0x1
1:806  0:000  Slots to scan [8]...
1:807  0:000  SPD[0]: Type 11 @0x50
1:825  0:018  XMP Profile1: 9*1/8ns
1:825  0:000  Found module with XMP version 1.3
1:825  0:000  Using XMP Profile1 instead of standard frequency 1777MHz
1:825  0:000  DDR speed 1777MHz
1:825  0:000  Slot: 0 Type 24 4096MB 1777MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:826  0:000  SPD[1]: Type 11 @0x51
1:844  0:018  XMP Profile1: 9*1/8ns
1:844  0:000  Found module with XMP version 1.3
1:844  0:000  Using XMP Profile1 instead of standard frequency 1777MHz
1:844  0:000  DDR speed 1777MHz
1:844  0:000  Slot: 1 Type 24 4096MB 1777MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:844  0:000  SPD[2]: Type 11 @0x52
1:863  0:018  XMP Profile1: 15*1/14ns
1:863  0:000  Found module with XMP version 1.2
1:863  0:000  Using XMP Profile1 instead of standard frequency 1866MHz
1:863  0:000  DDR speed 1866MHz
1:863  0:000  Slot: 2 Type 24 4096MB 1866MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:863  0:000  SPD[3]: Type 11 @0x53
1:881  0:018  XMP Profile1: 15*1/14ns
1:881  0:000  Found module with XMP version 1.2
1:881  0:000  Using XMP Profile1 instead of standard frequency 1866MHz
1:881  0:000  DDR speed 1866MHz
1:881  0:000  Slot: 3 Type 24 4096MB 1866MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:882  0:000  SPD[7]: Type 90 @0x57
Link to comment
Share on other sites
despeinao
despeinao
Posted
Posted

Hi all

 

Is it at all possible to boot Clover and macOS from an LSI SAS/SATA adapter, on a system without UEFI (using Legacy Boot)? 

 

I'm running 10.13.3 with latest Clover (4411).  I'm installed on an SSD so I am using APFS.  My motherboard is Gigabyte GA-X58A-UD3R, which does not have UEFI.

 

My LSI card is an LSI 9201-8i, which uses the 2008 controller chip and has latest firmware (v20).  I have this working in my macOS 10.13.3 installation using Astek drivers installed in EFI/CLOVER/kexts/Other.  Once booted, I can see and use drives attached to this controller.

 

But if possible I would like to move my main SSD to this controller and boot from it, so that I'm not using onboard SATA at all.  I've tried two things:

 

1. I attached my macOS/Clover SSD to the LSI and set it first in the boot order.  Clover started to load (I saw the cursor at top left, and then the number 6 briefly appears), but it then went to the "BIOS-like" interface where Continue is the first option, which the Clover Wiki describes as "This means that the boot file was successfully loaded and is working but it cannot find the file CloverX64.efi."

 

2. I also tried a new Clover install onto a USB drive.  This loads Clover OK, but drives installed on the LSI don't appear in the selection.

 

I suppose this is probably predictable - Clover can't read from the LSI controller without drivers, which don't exist.  However, I know other people are booting from RAID arrays on LSI MegaRAID cards.   But perhaps they can only do that if they have EFI support?  (Although the posts I've seen mentioning RAID boot are a year or two old, back on 10.11 and 10.12 and before APFS, and maybe that's changing something.)

 

Also I am able to boot other OS' direct from the LSI - Windows, Linux, etc.  Windows I have confirmed I can fully boot to from the LSI, even though I originally installed it on the onboard SATA.

 

But I suppose Clover and macOS must be a bit different, perhaps because of the need to fake EFI.  I'm sorry that I don't really understand the technical details of these boot loaders yet, though I am trying to learn more.  But I can imagine the Clover bootloader must be more complex than the basic one used to launch Windows, or Grub in Linux etc.

 

Is there any way to get this to work?  I suppose worst case I could have the main OS installed on a small SSD on my onboard SATA, then move /Users to the main SSD on the LSI.  But it'd be nice if I could just use the LSI controller and not need to split my macOS install over two SSDs.

 

Thanks in advance.

 

PS. in case it's of any importance, I've noticed that since I installed the LSI controller, I always see the number '6' in the top left of the screen just before the Clover GUI appears.  I saw in the FAQ that a '7' means problems, but I don't know what '6' means.  This happens regardless of how I boot Clover - from onboard SATA; from USB; and from a (failed) attempt to boot direct from the LSI.

 

You need to patch Clover so it reads the EFI bios from your adapter while booting in legacy mode. I opened a ticket with a patch for that, but it was rejected because it partially reverted an ancient patch and nobody knew if that might have ill effects to other users. Anyway, it works for me and probably for you too, as your motherboard is similar to mine. The ticket is here, and you can try replacing your boot file with the one attached. WARNING: Make sure that you have a way to boot from an USB in case this does not work for you!!!

 

EDIT: For this to work, you need to flash your adapter with _both_ the legacy mptsas2.rom and the UEFI x64sas2.rom

boot.zip

  • Like 3
Link to comment
Share on other sites
TheBloke
TheBloke
Posted
Posted

You need to patch Clover so it reads the EFI bios from your adapter while booting in legacy mode. I opened a ticket with a patch for that, but it was rejected because it partially reverted an ancient patch and nobody knew if that might have ill effects to other users. Anyway, it works for me and probably for you too, as your motherboard is similar to mine. The ticket is here, and you can try replacing your boot file with the one attached. WARNING: Make sure that you have a way to boot from an USB in case this does not work for you!!!

 

Wow that's awesome, thanks so much despeinao.  You are the wizard of LSI cards :)

 

One final question:  should I have Clover in SATA mode or BootBlockIO mode?  Or does it not matter once I have replaced with your boot.zip?  (I don't know if these options result in different boot files or not.)

 

Thanks so much again.  I will look into building Clover from source in future so I can keep applying your patch each time I need a new version.  Such a shame it was rejected - I will post on the ticket and maybe someone will consider if it can at least be added as an option in the Installer, for those who need it.

Link to comment
Share on other sites
despeinao
despeinao
Posted
Posted

 

One final question:  should I have Clover in SATA mode or BootBlockIO mode?  Or does it not matter once I have replaced with your boot.zip?  (I don't know if these options result in different boot files or not.)

 

 

This is just an installer option that places one boot file or another, so it does not matter as you are replacing the boot file. Note that if you reinstall or update Clover, the boot file will be overwritten.

  • Like 1
Link to comment
Share on other sites
TheBloke
TheBloke
Posted
Posted

You need to patch Clover so it reads the EFI bios from your adapter while booting in legacy mode. I opened a ticket with a patch for that, but it was rejected because it partially reverted an ancient patch and nobody knew if that might have ill effects to other users. Anyway, it works for me and probably for you too, as your motherboard is similar to mine. The ticket is here, and you can try replacing your boot file with the one attached. WARNING: Make sure that you have a way to boot from an USB in case this does not work for you!!!

 

EDIT: For this to work, you need to flash your adapter with _both_ the legacy mptsas2.rom and the UEFI x64sas2.rom

 

Just to confirm for the thread that this works :)  This is so awesome, thank you despeinao.

 

As stated in the edit, I needed to install the LSI card's x64sas2.rom on top of the existing mptsas2.rom.  The x64sas2.rom file can be found here for adapters with the 2008 chip.  It is the first download, under heading BIOS, called "UEFI_BSD_P20".  Make sure you've already flashed to P20 firmware and legacy BIOS (mpt2sas.rom) before applying this.  The same file for 2308 adapters should be available here.

 

I have added a comment to your patch, asking if it could be included as an optional file, in the same way that we have options for SATA/BiosBlockIO/BiosBlockIO for NVidia.   That way it could be easily used by those of us who need it, but would not be a risk to any normal user.

Link to comment
Share on other sites
Matgen84
Matgen84
Posted
Posted

 Any ideas?

 

 

Hi,

 

I found that in preboot.log: I don't know why same DDR use different XMP version and Frequency?

=== [ ScanSPD ] ===========================================
1:806  0:000  SMBus device : 8086 1E22 class=0C0500 status=Success
1:806  0:000  SMBus CmdReg: 0x3
1:806  0:000  Scanning SMBus [8086:1E22], mmio: 0xF7135004, ioport: 0xF000, hostc: 0x1
1:806  0:000  Slots to scan [8]...
1:807  0:000  SPD[0]: Type 11 @0x50
1:825  0:018  XMP Profile1: 9*1/8ns
1:825  0:000  Found module with XMP version 1.3
1:825  0:000  Using XMP Profile1 instead of standard frequency 1777MHz
1:825  0:000  DDR speed 1777MHz
1:825  0:000  Slot: 0 Type 24 4096MB 1777MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:826  0:000  SPD[1]: Type 11 @0x51
1:844  0:018  XMP Profile1: 9*1/8ns
1:844  0:000  Found module with XMP version 1.3
1:844  0:000  Using XMP Profile1 instead of standard frequency 1777MHz
1:844  0:000  DDR speed 1777MHz
1:844  0:000  Slot: 1 Type 24 4096MB 1777MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:844  0:000  SPD[2]: Type 11 @0x52
1:863  0:018  XMP Profile1: 15*1/14ns
1:863  0:000  Found module with XMP version 1.2
1:863  0:000  Using XMP Profile1 instead of standard frequency 1866MHz
1:863  0:000  DDR speed 1866MHz
1:863  0:000  Slot: 2 Type 24 4096MB 1866MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:863  0:000  SPD[3]: Type 11 @0x53
1:881  0:018  XMP Profile1: 15*1/14ns
1:881  0:000  Found module with XMP version 1.2
1:881  0:000  Using XMP Profile1 instead of standard frequency 1866MHz
1:881  0:000  DDR speed 1866MHz
1:881  0:000  Slot: 3 Type 24 4096MB 1866MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:882  0:000  SPD[7]: Type 90 @0x57
Link to comment
Share on other sites
Funky frank
Funky frank
Posted
Posted

I still cannot believe in what {censored} High Sierra turned:  Re-tried it last days, and it lags like hell with my 1050ti. This rig works quite smooth on Sierra. They said there would be a drastic increase of gfx performance thanks to Metal 2... Where is that boost? I cannot find it. In fact it obviously is much slower and laggy. What is Apple doing??  And this seems to be the same story on real Macs using eGPUs, too.  Even APFS is so much slower than HFS+, on a SSD! But normal Apple users seem to be too dumb to realize anything, there are barely accurate benchmarks in the media.  Even self calling pro news will not do their job properly. Is this the end of high quality OSX?  I wish I never switched from 10.9, but incompatibility forced me to do so. And then there is absolutely no working bug reporting @ Apple and nVidia, too. I tried that a bunch of times. It is a waste of your time. My last hope relies on an included/native pascal driver which was approved by Apple in future macOS versions. 

  • Like 1
Link to comment
Share on other sites
apianti
apianti
Posted
Posted

 

Hi,

 

I found that in preboot.log: I don't know why same DDR use different XMP version and Frequency?

=== [ ScanSPD ] ===========================================
1:806  0:000  SMBus device : 8086 1E22 class=0C0500 status=Success
1:806  0:000  SMBus CmdReg: 0x3
1:806  0:000  Scanning SMBus [8086:1E22], mmio: 0xF7135004, ioport: 0xF000, hostc: 0x1
1:806  0:000  Slots to scan [8]...
1:807  0:000  SPD[0]: Type 11 @0x50
1:825  0:018  XMP Profile1: 9*1/8ns
1:825  0:000  Found module with XMP version 1.3
1:825  0:000  Using XMP Profile1 instead of standard frequency 1777MHz
1:825  0:000  DDR speed 1777MHz
1:825  0:000  Slot: 0 Type 24 4096MB 1777MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:826  0:000  SPD[1]: Type 11 @0x51
1:844  0:018  XMP Profile1: 9*1/8ns
1:844  0:000  Found module with XMP version 1.3
1:844  0:000  Using XMP Profile1 instead of standard frequency 1777MHz
1:844  0:000  DDR speed 1777MHz
1:844  0:000  Slot: 1 Type 24 4096MB 1777MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:844  0:000  SPD[2]: Type 11 @0x52
1:863  0:018  XMP Profile1: 15*1/14ns
1:863  0:000  Found module with XMP version 1.2
1:863  0:000  Using XMP Profile1 instead of standard frequency 1866MHz
1:863  0:000  DDR speed 1866MHz
1:863  0:000  Slot: 2 Type 24 4096MB 1866MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:863  0:000  SPD[3]: Type 11 @0x53
1:881  0:018  XMP Profile1: 15*1/14ns
1:881  0:000  Found module with XMP version 1.2
1:881  0:000  Using XMP Profile1 instead of standard frequency 1866MHz
1:881  0:000  DDR speed 1866MHz
1:881  0:000  Slot: 3 Type 24 4096MB 1866MHz Vendor=Corsair PartNo=CMZ8GX3M2A1866C9 SerialNo=0000000000000000
1:882  0:000  SPD[7]: Type 90 @0x57

 

 Any ideas?

 

This is inaccuracy of SMB reading.

 

No, that's all SPD information. It's because while you may have the same brand and model of memory, they are from different manufacturing runs. You can tell because two are XMP 1.3 and two are XMP 1.2. I'm guessing you bought two dual channel kits? Maybe at different times?

  • Like 1
Link to comment
Share on other sites
frankiee
frankiee
Posted
Posted

When using the latest AptioMemoryFix in verbose mode, it says the following:

 

AMF: only 248/256 slide values are usable! Booting may fail!

Valid slides: 8-255

 

However, it boots just fine - most of the time. In some cases it stops here though, missing the usually following "+++++++++++" line. But this usually only happens when booting from a stick.

 

So, what does that mean? Fragmented Memory? And how can this be fixed?

 

And, since I have severe wake problems - i.e. machine crashes so hard it only can be shut down by pressing power button - (but only since 10.13) - could this be related?

Link to comment
Share on other sites
Matgen84
Matgen84
Posted
Posted

No, that's all SPD information. It's because while you may have the same brand and model of memory, they are from different manufacturing runs. You can tell because two are XMP 1.3 and two are XMP 1.2. I'm guessing you bought two dual channel kits? Maybe at different times?

Yes, bought two channels kits at different times. But all are the same brand and model. There is a solution?

Link to comment
Share on other sites
Matgen84
Matgen84
Posted
Posted

Hi Matgen84

Have you forced from bios in dram settings 1866 fequency without using auto xmp detection?

I don't remember. But since now, I don't have any problem with clover.  

 

I look at my parameters and I tell you

 

Thanks

 

Edit 1: same issue with XMP detection actived and Frequency memory auto

Link to comment
Share on other sites
TheBloke
TheBloke
Posted
Posted

 

Wow thanks for posting this.  Like Funky Frank I have been disappointed with GFX.  I just came back to macOS after ~9 months away, and I upgraded straight to 10.13.3 from 10.11.6, and since then I couldn't understand why simple things like fast scrolling a Chrome window, or even just playing a 1080P YouTube video seemed to have stuttering on my 980Ti.  I thought maybe it was system incompatibility or something, as I've been adding new HW and messing with DSDTs and lots of other changes.

 

I ran the script and it downgraded my drivers from 378.10.10.10.25.156 to .106.

 

I re-ran GeekBench 4 Compute benchmarks after and I did get slightly higher scores - just 1% higher, so within margin of error really.   But the main proof will be in normal usage, like videos and scrolling Chrome.  So far it's seemed smooth enough, including on a YouTube video which was where I first noticed some skipped frames.  But time will tell for sure.

 

Seems incredible that NVidia macOS drivers - or macOS itself - could be so bad that it's necessary to re-install much older drivers just to avoid choppiness on the Desktop.  But if it works, I'll take it!

 

Thanks again.

  • Like 2
Link to comment
Share on other sites
Trung_Nguyen
Trung_Nguyen
Posted
Posted

Sourceforge is migrating servers so it will be down intermittently.

 

I think you may be misguided. I am unsure what you are doing or trying to do but it doesn't really make a lot of sense. Or at least your explanation of it.....

Protocol's ReadKeyState will take forever while you hold the combination!

Also GRUB's KeyStatus can fetch the state of modifier keys, can it be used?

Link to comment
Share on other sites
mhaeuser
mhaeuser
Posted
Posted

Protocol's ReadKeyState will take forever while you hold the combination!

Also GRUB's KeyStatus can fetch the state of modifier keys, can it be used?

 

Kinda with apianti, you don't seem to understand the protocol, like, at all. It's not a key queue, it reports what is currently pressed, so, of course calling it in a loop would continuously return all the keys. That is its point.

Furthermore, it returns the Modifiers in the ModifierMap output variable. I have no clue what you want to do with GRUB's whatever functions, how that would help you or what your problem is in the first place.

 

The source of the driver is open in Clover, please just read it...

  • Like 1
Link to comment
Share on other sites
Trung_Nguyen
Trung_Nguyen
Posted
Posted

Kinda with apianti, you don't seem to understand the protocol, like, at all. It's not a key queue, it reports what is currently pressed, so, of course calling it in a loop would continuously return all the keys. That is its point.

Furthermore, it returns the Modifiers in the ModifierMap output variable. I have no clue what you want to do with GRUB's whatever functions, how that would help you or what your problem is in the first place.

 

The source of the driver is open in Clover, please just read it...

I read it. I didn't calling it in a loop. Just once in Refitmain?

main.c.zip

EDIT: Wait a min, so it write asynchronously to the array which is point in the parameter?

I thought it was the immediate result. Sorry

Link to comment
Share on other sites
mhaeuser
mhaeuser
Posted
Posted

I read it. I didn't calling it in a loop. Just once in Refitmain?

attachicon.gifmain.c.zip

EDIT: Wait a min, so it write asynchronously to the array which is point in the parameter?

I thought it was the immediate result. Sorry

It is an immediate result, though how do you expect the callee, which explicitely states the buffer is allocated by the caller in the header, to update a NULL-pointer you pass in? The caller must alloc the buffer. When you try to print stuff, you dereference NULL.

Link to comment
Share on other sites
 Share
Go to topic listing
×
×
  • Create New...